This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A 13-year-old girl in Kent, UK, committed suicide after social media recommendation algorithms repeatedly fed her disturbing videos that normalized self-harm.
Before her death on October 30, 2025, 13-year-old Blake Gallier reportedly viewed and reposted social-media videos normalizing suicide. After examining her phone, her family alleged that recommendation algorithms had repeatedly surfaced similar material. The platform or platforms involved have not been publicly identified, and an inquest into her death remains pending.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI that exposes users to harmful, abusive, unsafe or inappropriate content. May involve providing advice or encouraging action. Examples of toxic content include hate speech, violence, extremism, illegal acts, or child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.
Also classified as
The reports also mention a colleague who used deepfake technology to create non-consensual sexual imagery of four individuals for targeted harassment.
The primary incident involves an AI recommendation algorithm exposing a vulnerable minor to toxic, harmful, and suicide-normalizing content, leading to self-harm.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The harm was caused by the post-deployment operation of social media recommendation algorithms, which unintentionally created a harmful feedback loop of suicide-normalizing content for a minor.
Added categories peak at 3: Substantial (Psychological Harm, Direct)
This incident involves a tragic suicide linked to autonomous social media recommendation algorithms and a separate case of deepfake harassment. While these events highlight critical issues in online safety, child protection, and digital regulation, they represent localized societal harms and regulatory challenges rather than direct threats to national security.
The recommendation algorithms operate independently, responding to user interactions (such as liking a video) using learned patterns to autonomously serve content without human intervention.
Unacceptable Risk: The report describes an AI system used for behavioral influence leading to severe harm, specifically exploiting the psychological vulnerability of a minor to recommend self-harm content.
The primary function of the algorithms was to recommend content to the user based on her past interactions, creating a feedback loop of disturbing videos.
The reports provide clear details about the tragic outcome and the general role of social media algorithms. However, they do not name the specific social media platforms or technical details of the algorithms involved in Blake's case, and the deepfake stalking incident is only briefly mentioned in a news summary.
The suicide could be interpreted as primarily driven by underlying mental health issues, though the family strongly attributes the sudden shift to the algorithmic content loop.
Technology
Algorithmic Feedback Loop
Algorithm repeatedly fed disturbing content after user interacted with it once.
Inadequate Content Filtering
Platform algorithms failed to block videos normalising suicide and self-harm.
Data Inputs
User Engagement Data
Single likes or views triggered a stream of highly distressing related videos.
Worrying User Reposts
The user reposted worrying content without system intervention or alerting.
Human Factors
Lack of Parental Awareness
Mother was unaware of the disturbing content the child viewed and reposted.
Underestimating Social Media Risks
Family assumed the platforms were safe and did not research the content risks.
Youth Vulnerability
Young teens lack capacity to comprehend consequences of suicide normalisation.
Process and Methods
Deficient Age Verification
Platform allowed a thirteen year old to access and consume harmful content.
Inadequate Safety Interventions
No active mechanisms to detect and halt loops of self-harm content delivery.
Management
Prioritising Engagement Over Safety
Platform design prioritised keeping users online over shielding them from harm.
Regulatory Environment
Absence of Under-16 Bans
Lack of strict regulations banning vulnerable children under 16 from platforms.
Weak Tech Accountability Laws
Lack of legal requirements for platforms to preserve digital data after death.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
